Update documentation of the interface

Author: devmotion

README.md

@@ -56,7 +56,10 @@ use a custom distribution type `GaussianPrior`, the following methods should be
 implemented:
 ```julia
 # state that the distribution is actually Gaussian
-EllipticalSliceSampling.isnormal(::GaussianPrior) = true
+EllipticalSliceSampling.isnormal(::Type{<:GaussianPrior}) = true
+
+# define the mean of the distribution
+Statistics.mean(dist::GaussianPrior) = ...
 
 # define how to sample from the distribution
 # only one of the following methods is needed:
@@ -68,6 +71,9 @@ Base.rand!(rng::AbstractRNG, dist::GaussianPrior, sample) = ...
 
 # specify the type of a sample from the distribution
 Base.eltype(::Type{<:GaussianPrior}) = ...
+
+# in the case of mutable samples, specify the array size of the samples
+Base.size(dist::GaussianPrior) = ...
 ```
 
 ### Log likelihood

src/types.jl

@@ -8,7 +8,7 @@ struct Model{P,L}
     loglikelihood::L
 
     function Model{P,L}(prior::P, loglikelihood::L) where {P,L}
-        isnormal(prior) ||
+        isnormal(P) ||
             error("prior distribution has to be a normal distribution")
 
         new{P,L}(prior, loglikelihood)

src/utils.jl

@@ -21,5 +21,5 @@ end
 Test whether a distribution is normal.
 """
 isnormal(dist) = false
-isnormal(::Normal) = true
-isnormal(::AbstractMvNormal) = true
+isnormal(::Type{<:Normal}) = true
+isnormal(::Type{<:AbstractMvNormal}) = true